Ask HN: What's your containerized LLM coding stack?

Hi HN,

As many of you, I've switched to Cursor for code-editing over the past year and am very curious to switch to coding agents pushing PRs to my repositories.

However, I'm not sure there's some consensus as to which stack use. What's the best way to containerize these agents?

I have several servers running at home, ideally I'd like a few running on each of them.